Introduction
Even before the use of RFID among supply chain companies, there had been a growing need for a standardized and fully automated, continuous means of exchanging information about items and transactions that occur among trading partners (e.g. manufacturers, wholesalers and retailers). Today, this need is gradually being satisfied through "data synchronization," a technology that compares and reconciles selected transactions and other data from partners to ensure that each trading partner has the same information about items of interest at the same point in time. It is the Internet that provides the communications medium that is used for this data exchange.

Several studies have shown that when supply-chain partners have automatic access to each other's selected item and transaction information, it produces a manifold increase in efficiency and a corresponding reduction of errors. According to a study by Yankee Group, improved data synchronization combined with RFID technology has the potential to cut between $2 billion and $4 billion per year in costs from the consumer packaged goods and retail industries.

With the growing use of RFID in the supply chain sector and its ability to instantly and automatically record inventory movements (as well as identify each item uniquely), the need for data standardization and synchronization has become particularly crucial. This is because of two key factors: 1) products can now be automatically identified at various points in the supply chain, regardless of the visibility of the tag, which radically speeds the identification of items, cases and pallets as these move between inventory spaces; 2) RFID allows the identification of each unique item, not just each type of item. In other words, companies have the ability to identify each unique can of soda, package of CDs, bottle of medicine, etc., and not just the brand, type and unit of measure as is typically done with bar-codes. This item-specific data allows close tracking of the date of manufacture, the movements of the item, and much more.

When all entities in the supply chain are automatically sharing standardized data (data synchronization), tremendous increases in automation can be achieved at many levels of the supply-chain. In addition, partners located in any part of the world will be able to more easily and efficiently do business with each other. Without clean and fully standardized data, however, RFID will likely enable faster exchange of potentially incorrect and/or incomplete information.

Data synchronization typically occurs over subscription-based information exchange networks. Among supply-chain companies, there are two primary data synchronization networks. The first is the Global Data Synchronization Network (GDSN), which allows companies to automatically exchange detailed static product information; in other words, everything that remains the same about a particular product; e.g., manufacturer, brand, unit of measure, weight, type, style, etc. The second is the EPCglobal Network™, which is mostly concerned with dynamic information about products; i.e., transit times, manufacture dates, and each warehouse, truck and store through which the product traveled on its way from the manufacturer to the end user. It is the EPCglobal Network that particularly benefits from the use of RFID technology. More about the GDSN and the EPCglobal Network shortly.

As previously mentioned, data synchronization occurs over the Internet, however, it normally does this without the need for people to use Web browsers and locate or send data. When data synchronization is working correctly, the locating, sending and retrieving of necessary data occurs automatically. If you are familiar with how Web services work, data synchronization operates much within the same model.

GTIN and EPC™ Codes
Before a company can begin to use either of the previously mentioned supply-chain data synchronization networks, it must first ensure that its products are associated with a proper Global Trade Identification Number (GTIN) and Electronic Product Code (EPC). Nearly every product sold in the United States has a UPC number (that is translated into a bar code), which qualifies as a GTIN, so that part is already handled for most companies. The next part is to obtain an EPC code (a code that is fundamentally based on the GTIN) for each product from EPCglobal (www.epcglobalinc.org). It is the EPC code that is encoded on RFID tags. EPC codes identify the GTIN (and other information), and can optionally specify a serial number for each iteration of the product (unlike a UPC code). As the product moves through the supply chain, EPC codes allow the recording of dynamic product information that is specific to each individual item, such as expiration dates as well as shipping, handling and storage details.

Global Data Synchronization Network (GDSN)
GDSN is being used by many companies today that aren't necessarily RFID-enabled. As already mentioned, this is because the data that flows over this network contains only static information about products;in other words, everything that remains the same about a particular product. GDSN provides many benefits for entities along the supply chain; for instance, if a manufacturer has a new product or if the characteristics of an existing product change, every retailer in the world that is connected through this network will have the means to instantly know about this information (that is if they have chosen to "subscribe" to this data). Also, if a retailer wants to know about every product (or manufacturer) available on the global market with particular attributes, it can rapidly gain a complete listing with extensive information. Once this kind of automated exchange of data is established, a retailer could—if they chose to—automatically generate reorders over this network based on its inventory reaching predefined levels.

Data pool — These are repositories of data that describe each company's products. This data is expressed in a generic language (typically a derivative of extensible markup language [XML]) so that other companies can easily use this data. Since the information on a bar-code or on an RFID chip is typically only the GTIN (expressed as either a UPC for bar-codes or an EPC for the RFID tags), it is the data pool that contains extensive information about the item. When a supply chain entity wants to use data synchronization, it must subscribe to a data pool organization and have its item and company information hosted there. For a supply chain entity, it is the data pool that acts as the point of entry to the GDSN. Data pool organizations include: UCCnet, Transora, Worldwide Retail Exchange, and others.

GS1 Global Registry™ — This is a global data registry that hosts all of the "lookup" information for all of the entities that are part of the GDSN in order to identify the data pool in which the requested information can be found. The GS1 Global Registry is like the domain name server (DNS) registries that cross-reference all World Wide Web (WWW) addresses to their hosts so that when a person types a URL in a browser (i.e., www.tlashford.com), the registry looks up the server where the web site is hosted and your browser is automatically redirected to it.

Based on these definitions, the steps in the Global Data Synchronization Network can be described as follows:

Step 1. Load data. A supplier sends detailed information about its GTIN registered items to its data pool (data pools currently support over 150 standardized product and company attributes). This data is described in a standardized format, typically using an XML schema called Product Markup Language (PML).

Step 2. Register data. This product information is registered with the GS1 Global Registry which acts as an index that lets other members of GDSN (regardless of their data pool) locate item information from any other entity.

Step 3. Subscription request. A retailer, subscribed to its own data pool partner (recipient data pool), sends a request to locate products with particular specifications. This is routed to the GS1 Global Registry, which does a lookup and then routes the request to the appropriate source data pool.

Step 4. Publish data. Once the appropriate source data pool is contacted and the identity of the requester is verified, the data is published from the source data pool to the recipient data pool, and finally routed to the retailer and its internal backend systems.

Keep in mind that most of these steps are automated, which allow trading partners to instantly know when a new item has been made available, an existing item has been changed, an order has been placed, or an order has been shipped, etc.

EPCglobal Network™

The EPCglobal Network is a virtual network that was developed according to standards from the EPCglobal organization, which takes supply chain data synchronization a giant step beyond GDSN. That's because the EPCglobal Network harnesses the advantages of RFID to exchange dynamic information about products. As mentioned earlier, this relates to events occurring with individual items, not just the passing of generalized product information. For instance, the EPCglobal Network along with RFID technology can track (via the transmission of the EPC by RFID technology) transit times, manufacture dates, and even each warehouse, truck and store through which the product traveled on its way from the manufacturer to the end user. Eventually, "smart shelves" equipped with RFID readers will be able to detect when inventory quantities on store shelves reach a predefined level which could automatically trigger a reorder from a vendor.

Besides the kind of data that is synchronized, one of the main differences between the GDSN and EPCglobal Networks is that with GDSN the data pool is hosted by a third party. The EPCglobal Network, on the other hand requires that each company be its own data pool. In other words each company's own dynamic data is hosted by their own Web server. One of the reasons for this is because when it comes to dynamic data, it is very transaction intensive, so porting this data first to a third-party data pool would affect the ability to quickly access this information.

Definitions

EPCIS (EPC information systems). The EPCIS is essentially the data pool for dynamic transaction information, although it resides on each company's own web server instead of on a third party's server. This server records dynamic data about the movement (and other variables) of each inventory item.

ONS (Object Name Server). This is a registry, residing on the Internet, that provides a function similar to the GS1 Global Registry in the GDSN data synchronization model. It hosts all of the "lookup" information for all of the entities that are part of the EPCglobal Network in order to identify the trading partner's EPCIS from which the appropriate transaction information can be found.

Based on these definitions, the steps in the EPCglobal Network can be described as follows:

Step 1. Load data. A supplier scans RFID tags as inventory is moved in and out of inventory spaces. This data that contains information about the movement of this inventory is automatically loaded to its EPCIS (and its own internal applications) via EPC middleware. As with GDSN, this data is described in a standardized format, typically using an XML schema called Product Markup Language (PML).

Step 2. Register data. The supplier's EPCIS is registered with the ONS (and any other appropriate directory services) so that the supplier can be identified to the appropriate trading partners in order to automatically access the transaction information that describes the movement (or other dynamic data) of the inventory item.

Step 3. Subscription request. A retailer sends a request to locate information from a selected supplier about transactions concerning specific items. This request is routed to the ONS, which does a lookup and then routes the request to the appropriate EPCIS of the trading partner.

Step 4. Publish data. Once the appropriate EPCIS is contacted and the identity of the requester is verified, the data is "published" from the source EPCIS to the recipient EPCIS, and finally routed to the requester's internal systems.